adversarial-check
Use when the model commits to the first plausible explanation or conclusion and then continues strengthening it without seriously exploring counterexamples…
Skill by Bogyie
Chalk Labs
Need help deploying adversarial-check?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/bogyie-adversarial-check
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/adversarial-check
What's inside
- Skill name
adversarial-check- Description
- Use when the model commits to the first plausible explanation or conclusion and then continues strengthening it without seriously exploring counterexamples, competing hypotheses, or disconfirming evidence.
- License
- Apache-2.0
- Instructions
- 19 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# adversarial-check Do not use the first plausible explanation as the final conclusion. Identify the leading hypothesis A, then perform these steps separately: 1. Find the strongest evidence supporting A. 2. Find the strongest evidence that A is wrong. 3. Construct the strongest competing hypotheses. 4. Identify facts each competing hypothesis fails to explain. 5. Identify observations or evidence that would reverse the current conclusion. Do not add a token counterargument merely for balance. Assume the current conclusion is wrong and construct the most credible alternative explanation available. Compare the hypotheses using the same criteria before deciding on the final conclusion and confidence level.
From Bogyie/llm-reliability-skill, Apache-2.0.
More from Bogyie/llm-reliability-skill
- 0.524score
orthogonal-reviewer
Bogyie
Use when multiple reviewers or models are likely to share the same blind spots because they see the same context, use the same evaluation criteria, or simply…
SkillBFreshness: BClaude Code +9 - 0.524score
process-integrity
Bogyie
Use when the final result may be correct but the agent could violate process constraints along the way, such as modifying forbidden files, skipping approvals…
SkillBFreshness: BClaude Code +9 - 0.524score
replan-gate
Bogyie
Use when an agent keeps following its initial workflow, such as inspect -> edit -> test, until a hard failure occurs, even when new evidence indicates it…
SkillBFreshness: BClaude Code +9 - 0.524score
change-impact-analysis
Bogyie
Use when a coding agent focuses narrowly on resolving the current issue or implementing the requested feature and may miss callers, consumers, public APIs…
SkillBFreshness: BClaude Code +9 - 0.524score
constraint-enforcer
Bogyie
Use when the model knows and can restate constraints such as forbidden files, security rules, compatibility requirements, or process rules, but may still…
SkillBFreshness: BClaude Code +9 - 0.524score
source-arbitrator
Bogyie
Use when official documentation, source code, release notes, GitHub issues, blogs, community posts, search results, and model memory conflict and must be…
SkillBFreshness: BClaude Code +9 - 0.524score
docs-before-code
Bogyie
Use when code appears easy enough to modify directly and the agent is likely to skip repository instructions, README files, ADRs, architecture documents…
SkillBFreshness: BClaude Code +9 - 0.524score
evidence-first-reporting
Bogyie
Use when the model produces a persuasive reasoning narrative that blurs the boundary between observed facts, inferred conclusions, assumptions, tool results…
SkillBFreshness: BClaude Code +9 - 0.524score
spec-over-tests
Bogyie
Use when a coding agent is likely to optimize for visible tests, hard-code test cases, weaken tests, or treat passing the current test suite as equivalent to…
SkillBFreshness: BClaude Code +9 - 0.524score
independent-review
Bogyie
Use when the model is reviewing its own previous output and is likely to defend its earlier reasoning, overlook self-generated errors, or treat its previous…
SkillBFreshness: BClaude Code +9 - 0.524score
context-distiller
Bogyie
Use when the necessary information already exists in a long context, large repository, document set, or search result set, but important constraints, design…
SkillBFreshness: BClaude Code +9 - 0.519score
terminology-discipline
Bogyie
Use when the model assigns special meanings to ordinary words, invents local pseudo-terminology, or repeatedly uses expressions that are not established in…
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Aug 14, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/Bogyie/llm-reliability-skill
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from Bogyie
- 0.524score
docs-before-code
Bogyie
Use when code appears easy enough to modify directly and the agent is likely to skip repository instructions, README files, ADRs, architecture documents…
SkillBFreshness: BClaude Code +9 - 0.524score
evidence-first-reporting
Bogyie
Use when the model produces a persuasive reasoning narrative that blurs the boundary between observed facts, inferred conclusions, assumptions, tool results…
SkillBFreshness: BClaude Code +9 - 0.524score
context-distiller
Bogyie
Use when the necessary information already exists in a long context, large repository, document set, or search result set, but important constraints, design…
SkillBFreshness: BClaude Code +9 - 0.524score
change-impact-analysis
Bogyie
Use when a coding agent focuses narrowly on resolving the current issue or implementing the requested feature and may miss callers, consumers, public APIs…
SkillBFreshness: BClaude Code +9 - 0.524score
constraint-enforcer
Bogyie
Use when the model knows and can restate constraints such as forbidden files, security rules, compatibility requirements, or process rules, but may still…
SkillBFreshness: BClaude Code +9 - 0.524score
independent-review
Bogyie
Use when the model is reviewing its own previous output and is likely to defend its earlier reasoning, overlook self-generated errors, or treat its previous…
SkillBFreshness: BClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get adversarial-check running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
adversarial-check skill. Agent Mart. https://agentmart.dev/skills/bogyie-adversarial-check (last updated 2026-10-24)